Author:
LIU Mingyang,YANG Qiming,HU Guanhua,GUO Yan,ZHANG Jiandong
Abstract
In response to the difficulty in deploying anchor box based methods in 3D object detection due to the increase in spatial dimensions, this paper studies a point cloud object detection algorithm based on set prediction. This article proposes a Transformer based 3D point cloud object detection algorithm, and combines the characteristics of point clouds in autonomous driving scenarios to propose an improved spatial modulation attention and heat map initialization strategy for training acceleration and query initialization, achieving good detection performance in shallow networks. This article compares it with other algorithms on the KITTI dataset, and the results show that our algorithm has reached an advanced level in performance. We also conducted ablation experiments on the main components of the algorithm to verify the contribution of each module to the detection effect.
Reference17 articles.
1. A novel topology framework for simultaneous topology, size and shape optimization of trusses under static, free vibration and transient behavior
2. VASWANI A, SHAZEER N, PARMAR N, et al. Attention is all you need[C]//31st International Conference on Neural Information Processing Systems, New York, 2017: 6000–6010
3. KIRILLOV A, USUNIER N, CARION N, et al. End-to-end object detection with transformers[C]//2020 European Conference on Computer Vision, Cham, 2020: 213–229
4. ZHOU Quan, NI Yinghao, MO Yuwei, et al. FMA-DETR: a Transformer object detection method without encoder[J/OL]. (2023-10-16)[2023-11-30]. [Article] (in Chinese)